多尺度耦合经验正交分解法反演海洋次表层温盐结构  

Inversion of ocean subsurface temperature and salinity structure by multiscale coupled empirical orthogonal decomposition

作　　者：张明琪 徐永生[2,3,4] 杨树国 ZHANG Mingqi;XU Yongsheng;YANG Shuguo(Qingdao University of Science and Technology,Qingdao 266061,China;Institute of Oceanography,Chinese Academy of Sciences,Qingdao 266071,China;Laoshan Laboratory,Qingdao 266061,China;University of Chinese Academy of Sciences,Beijing 100094,China)

机构地区：[1]青岛科技大学,山东青岛266061 [2]中国科学院海洋研究所,山东青岛266071 [3]崂山实验室,山东青岛266061 [4]中国科学院大学,北京100094

出　　处：《海洋科学》2024年第3期1-12,共12页Marine Sciences

基　　金：崂山实验室科技创新项目(LSKJ202201406-2);NSFC-山东省联合基金项目(U1406401);国家自然科学基金项目(41906027);中国科学院战略性先导科技专项(XDB42000000)。

摘　　要：针对提升利用卫星观测海面信息重构三维温盐场精度的问题,本文提出了一种基于多尺度耦合正交分解的三维温盐场重构方法。该方法利用多尺度耦合正交分解对历史温盐剖面进行从大尺度到小尺度的迭代分解,以分层提取不同尺度三维温盐场的特征信息,然后分别建立海洋表面卫星观测与不同尺度的三维温盐场特征信息的重构模型,从而达到三维温盐场重构的目的。本文分别利用多尺度耦合正交分解法与单层正交分解法进行三维温盐场重构,结果显示,多尺度耦合正交分解优于单层正交分解法,且随着分解的层次不断细化,重构温盐场的精度及其垂直梯度精度均呈现明显的提升,其中组合4精度提升程度最高,分别提升了25.57%和27.58%;同时,对比HYCOM模式数据,多尺度耦合正交分解重构方法能有效地捕捉次表层海洋的空间特征信息。总体上,本文耦合经验正交分解法反演的温盐场与Argo温盐场偏差较小,反演精度较好,在空间分布上趋于一致。To address the problem of reconstructing the three-dimensional ocean thermohaline salinity fields based on sea surface information obtained through satellite observation,a three-dimensional thermohaline field–reconstruction method based on multiscale coupled empirical orthogonal decomposition is proposed in this paper.The proposed method involves using a multiscale coupled empirical orthogonal function to iteratively de-compose historical temperature and salinity profiles from a large to a small scale to extract the feature information of three-dimensional thermohaline and salinity fields at different scales in a hierarchical manner.Then,models of the collected satellite information are established and the characteristic information of the involved three-dimensional thermohaline field is obtained at different scales to achieve three-dimensional thermohaline field reconstruction.Herein,multiscale coupled empirical and single-layer orthogonal decompositions are used to recon-struct three-dimensional thermohaline and salinity fields.The results indicate that the former is superior to the latter.Furthermore,with continuous refinement of the decomposition levels,the accuracy of the reconstructed thermoha-line field and its vertical-gradient accuracy are significantly improved,with Scheme 4 resulting in the highest im-provements(25.57%and 27.58%,respectively).Meanwhile,compared with the HYCOM,the proposed method can more effectively capture the spatial characteristics of the involved subsurface ocean.In general,thermohaline-field inversion based on the proposed method deviates less from the Argo thermohaline field;moreover,the accuracy of the inversion is better.Further,the obtained spatial distrtbution tends to be consistent,aiding in effectively analyz-ing the three-dimensional structure and variation characteristics of the ocean thermohaline field.

关 键 词：卫星观测 多尺度耦合正交分解 温盐梯度 异常矩阵 

分 类 号：P731[天文地球—海洋科学]